Skip to content

Current state of the Cloud Pak Deployer🔗

The below picture indicates the current state of the Cloud Pak Deployer, which infrastructures are supported to provision or use OpenShift, the storage classes which can be controlled and the Cloud Paks with cartridges and components. Current state of the deployer

Support statement🔗

Cloud Pak Deployer has proven to be a useful tool to install OpenShift, the IBM Cloud Paks and watsonx through a unified and unattended installation script. It has been used by many IBM field specialists, IBM business partners, clients and even several development teams are using the tools. A question that often comes up: "Is Cloud Pak Deployer supported by IBM Support". The answer is NO. for the tool itself, similar to any bash, Python, Ansible, Chef, Terraform, CI/CD pipelines, or other scripts that would be developed by IBM Expert Labs, business partners or clients.

However, the IBM products which are deployed using the tool can continue to be supported, given that the steps executed for the installation correspond with the official product documentation. For example, the installation of Cloud Pak for Data and watsonx and its services are done through the apply-olm and apply-cr subcommands which are part of cpd-cli manage.

To create better insights into which steps have been executed for the installation of the products, Cloud Pak Deployer now generates a log file $STATUS_DIR/log/deployer-activities.log, which logs all the steps and which can be used to reproduce an installation using the standard cpd-cli and oc commands found in the documentation. Activity logging has currently been implemented only for Existing OpenShift installations of Cloud Pak for Data and watsonx. Other Cloud Paks will follow.

To follow the overall installation process from a single place, Cloud Pak Deployer maintains a unified installation progress log in $STATUS_DIR/log/installation-progress.log. The progress log lists all installation steps of the run (for example, 10 - Validate through 80 - Smoke tests) with their status (PENDING, RUNNING, COMPLETED, FAILED or SKIPPED), their start and completion timestamps, the duration of each step, and the log file or directory where the detailed logs for the step can be found. The file is updated while the installation is running, so it can be checked at any time without having to follow the individual component logs. When a step fails, the progress log identifies the failed step and points at cloud-pak-deployer.log where the detailed error information is recorded. The progress log of a previous run is archived together with cloud-pak-deployer.log. A step that is still shown as RUNNING after an interrupted run indicates that the deployer was terminated while that step was executing.

Additionally, you can use Cloud Pak Deployer to generate an activity log only using the --dry-run command-line option, run as an OpenShift user with cluster-reader role. This will be convenient if you do not want to let deployer do the installation but generate the steps you can then execute manually or through a script.

For the "general" support statement, check the Apache license which is referenced in the GitHub repository.